Weakened Flexural Strength of Nanocrystalline Nanoporous Gold by

Mar 16, 2016 - We prepared three gold–silver precursor alloys, well-annealed, prestrained, and high-energy ball-milled, from which nanoporous gold s...
0 downloads 6 Views 5MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Weakened Flexural Strength of Nanocrystalline Nanoporous Gold by Grain Refinement Eun-Ji Gwak† and Ju-Young Kim*,†,‡ †

School of Materials Science and Engineering, UNIST (Ulsan National Institute of Science and Technology), Ulsan 44919, Republic of Korea ‡ KIST-UNIST Ulsan Center for Convergent Materials, UNIST, Ulsan 44919, Republic of Korea S Supporting Information *

ABSTRACT: High density of grain boundaries in solid materials generally leads to high strength because grain boundaries act as strong obstacles to dislocation activity. We find that the flexural strength of nanoporous gold of grain size 206 nm is 33.6% lower than that of grain size 238 μm. We prepared three gold−silver precursor alloys, well-annealed, prestrained, and high-energy ball-milled, from which nanoporous gold samples were obtained by the same free-corrosion dealloying process. Ligaments of the same size are formed regardless of precursor alloys, and microstructural aspects of precursor alloys such as crystallographic orientation and grain size is preserved in the dealloying process. While the nanoindentation hardness of three nanoporous golds is independent of microstructural variation, flexural strength of nanocrystalline nanoporous gold is significantly lower than that of nanoporous golds with much larger grain size. We investigate weakening mechanisms of grain boundaries in nanocrystalline nanoporous gold, leading to weakening of flexural strength. KEYWORDS: Nanoporous gold, grain boundary density, dislocation density, ligament size, nanoindentation, flexural strength of relative density and ligament size: σ* = Cs[σ0 + kL(−1/2)](ρ*/ ρs)(3/2), where σ* is the ligament yield strength, Cs is the fitting coefficient, σ0 is the yield strength of bulk Au, k is a Hall-Petchtype coefficient, L is ligament size, and ρ*/ρs is the relative density of np-Au.18 Farkas et al. observed tension-compression asymmetry in strength and Poisson’s ratio by atomistic simulations on np-Au for ligament size below 15 nm.19 As reported in previous research on the mechanical behavior of np-Au, ligament size is the primary microstructural factor determining strength (or hardness), which is in line with “smaller is stronger” phenomenon revealed by nanopillar compressions.20 The fundamental mechanism in the “smaller is stronger” phenomenon is that, with smaller dimensions, dislocations can move and escape at free surfaces before interacting with obstacles or other dislocations, resulting in dislocation starvation.20−24 Obstacles to dislocation motion, such as grain boundaries and other dislocations, are also important in determining mechanical behavior of materials in small sample size. Effects of grain boundary and initial dislocation density have been studied. Combined with reduced external sample size in nanopillars, Ng and Ngan suggested that dislocations accumulate at grain boundaries, resulting in increased strength and strain hardening in microsized aluminum (Al) bicrystal pillars.25 Meanwhile, Kunz et al. observed that grain boundaries

N

anoporous gold (np-Au) has received considerable attention in applications for catalysts,1−3 sensors,4,5 and actuators6−8 due to its comparatively easy fabrication9,10 and chemical stability as well as its open-cell porous structure with high surface-to-volume ratio. However, its fragility has been retarding practical applications,11−13 so that its mechanical behavior has been widely investigated.11−19 Li and Sieradzki first reported the mechanical properties of np-Au prepared by electrochemical dealloying and postheat treatment using threepoint bending tests. They reported transition of ductile and brittle fracture depending on ligament size.11 Volkert et al., investigating micron-sized np-Au column compression fabricated by focused ion beam, show that the strength of np-Au is not dependent on the column diameter but on the ligament size and that the ligament strength approaches the theoretical strength.14 Biener et al. also found that nanoindentation hardness and compressive strength of np-Au depend strongly on ligament size in a way not described by the conventional Gibson-Ashby scaling model15 taking into account relative density only.16 Lee et al. fabricated free-standing dog-boneshape np-Au samples by e-beam lithography and carried out deflective tensile tests using nanoindentation.13 Through the experiments, they measured elastic modulus, yield strength, and residual stress potentially formed by volume shrinkage during dealloying. Sun et al., using in situ nanoindentation on np-Au in TEM,17 observed that dislocations were generated in the outermost ligament layer and moved to nodes where they may interact with others coming from other ligaments. Hodge et al. suggested a modified scaling equation incorporating the effects © XXXX American Chemical Society

Received: January 6, 2016 Revised: March 4, 2016

A

DOI: 10.1021/acs.nanolett.6b00062 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Nanoindentation data were analyzed from 80 nanoindentations for each sample. Three-point bending tests were performed with a custom-build three-point bending jig of 4 mm span length at constant cross-head speed of 0.002 mm/s using a microscale universal testing machine (Instron 5948). Cuboid samples for the bending tests were machined from both-sidespolished precursor alloys to dimensions of 2 mm width × 8 mm length × about 0.5 mm thickness using a nanomachining instrument (FANUC Ultra-Precision Nano Machine). Span length and sample dimension for three-point bending tests were determined by a standard test method.29 Figure 1a−c shows EBSD inverse pole figure (IPF) maps for precursor alloys. Grain sizes were measured for more than 10

in Al bicrystal pillars act as dislocation sinks, leading to lower strain-hardening.26 Looking at initial dislocation density, Lee et al. found that high initial dislocation densities in Au nanopillars obtained by precompression soften the pillars by acting as dislocation sources operating at lower stress than in pristine Au nanopillars.27 Bei et al. found intermediate prestrained (4−8% compression) single-crystalline Mo alloy micropillars show stochastic and scattered features, while extreme prestained (11% compression) pillars behave like the bulk material.28 Although on the basis of previous research the mechanical behavior of np-Au could be affected by microstructural factors of grain boundary and initial dislocation density, these issues have not been fully studied. Here we fabricate np-Au samples with different grain boundary and initial dislocation density. We prepare wellannealed, prestrained, and high-energy ball-milled Au−Ag precursor alloys. Since the microstructures of precursor alloys such as crystallographic orientation and grain size are preserved during dealloying, we obtain nanocrystalline np-Au with grain size 206 nm from ball-milled precursor alloy. We find that the flexural strength of nanocrystalline np-Au is significantly lower than that of annealed and prestrained np-Au, while nanoindentation hardness of annealed, prestrained, and nanocrystalline np-Au is almost identical. The high grain boundary density in nanocrystalline np-Au weakens flexural strength. We investigate possible weakening mechanisms of grain boundaries in the flexural strength of np-Au. Np−Au samples were made by free corrosion dealloying of precursor Au−Ag alloys (Ag72Au28 in at. %), which were prepared from Au (99.99%) and Ag (99.99%) pellets by melting at 1100 °C and homogenizing for 72 h at 850 °C under a N2 atmosphere. From this homogenized state, three different “annealed”, “prestrained”, and “ball-milled” precursor alloys were prepared. Samples called “annealed” and “prestrained” were cut into about 1 mm-thick disks, polished on both sides with 0.25 μm diamond suspension, and annealed for 24 h at 850 °C to relieve stress. The as-annealed sample is referred to as “annealed” precursor alloy. Some annealed precursor alloys were compressed to 5% compressive engineering strain using a universal testing machine (Instron 5982) at loading rate 50 μm/min. This compressed sample is called the “prestrained” precursor alloy. The “ball-milled” sample was prepared by highenergy ball milling (SPEX Mixer Mill 8000D). After homogenization, pieces of this alloy were sealed in stainless steel vials together with stainless steel balls, and shaken at 1060 back-and-forth cycles/min for an hour. Ball-milled alloys were cut into about 1 mm-thick disk and polished gently on both sides with 0.25 μm diamond suspension. For free corrosion dealloying to obtain np-Au, these three precursor alloys were dipped in 35% nitric acid at 80 °C for 72 h and rinsed repeatedly with ethanol and D.I. water. Np−Au samples were examined by field emission scanning electron microscopy (FE-SEM, FEI NovaNano 230) for imaging, energy-dispersive X-ray spectroscopy (EDS) to measure amount of residual Ag, and electron backscatter diffraction (EBSD, TSL-OIM) to obtain an inverse pole figure map. Nanoindentations (Keysight G200) were carried out on precursor alloys and np-Au samples with a Berkovich indenter using continuous stiffness measurement (CSM) in XP module with force capacity 500 mN for the precursor alloys and dynamic contact module II (DCM II) with force capacity 30 mN for np-Au samples. Nanoindentations were performed to a maximum indentation depth of 2.5 μm with strain rate 0.05 s−1.

Figure 1. Microstructural change and formation of nanoporosity by dealloying for three samples; annealed, prestrained, and ball-milled npAu. Electron backscatter diffraction (EBSD) inverse pole figure maps (a−c) before and (d−i) after dealloying process showing preservation of grain structure during dealloying. Ligaments of (g) annealed and (h) prestrained np-Au have identical crystallographic orientation coming from same grain in precursor alloy. Ligaments of (i) ball-milled np-Au have high grain-boundary density. (j−l) SEM images of three np-Au samples with almost identical ligament size regardless of variation in initial dislocation density and grain size of precursor alloys.

EBSD IPF images of each sample by the intercept method. Grain sizes are 238 μm for the annealed precursor alloy, 266 μm for the prestrained one, and 206 nm for ball-milled one. Figure 1c shows that ultrafine grains are uniformly formed by the high energy ball-milling process. No evolution of texture was observed in the prestraining and ball-milling process. Figure 1d−i presents EBSD inverse pole figure maps for np-Au after dealloying for three precursor alloys, showing that grain morphology and crystallographic orientation were preserved in free corrosion dealloying for all three samples,30,31 suggesting that nanoporosity formation is accomplished by short-range B

DOI: 10.1021/acs.nanolett.6b00062 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters surface diffusion of Au atoms during dealloying.32,33 Figure 1j−l show SEM images for np-Au samples of the three different precursor alloys, annealed, prestrained, and ball-milled. Ligament sizes were determined from at least 100 measurements in SEM images on the basis of necks at the center of connects in ligaments, possibly the thinnest part in ligaments. Ligament sizes are almost identical: 73 (±19.5) nm for annealed, 70 (±21.9) nm for prestrained, and 75 (±24.2) nm for ball-milled np-Au. The prestrained precursor alloy is presumed to have higher initial dislocation density than the annealed one, and the ball-milled precursor alloy is most likely to have greater densities of initial dislocations and grain boundaries than the other two precursor alloys. Size and morphology of ligaments and pores are known to be controlled by precursor composition,34 dealloying conditions,35,36 and post-treatment.37 On the other hand, the almost identical ligament sizes for three np-Au samples imply that changes in initial dislocation density and grain size in a precursor alloy are unlikely to change the nanoporous structure obtained by dealloying. Figures 2a and b show typical and average nanoindentation force-depth curves for the three precursor alloys. Forces

deformation induced by the high-energy ball-milling process. The hardness of prestrained sample is 31.3% higher than that of annealed sample. This could be caused by an increase in initial dislocation density generated by compressive plastic deformation. Assuming that the prestrained and annealed samples have the same plastic constraint factor and obey the Taylor hardening model, the initial dislocation density of the prestrained precursor alloy is about 2 orders of magnitude greater than that of the annealed one (by comparison of macroscopic hardness in the Nix−Gao model;38 Supporting Information, Figure S1). This value agrees well with that evaluated from the change in flow and yield stress in the compression stress−strain curve up to 5% in the prestraining process. Figure 2c and d shows typical and average nanoindentation force-depth curves for np-Au made from three precursor alloys; the curves are almost identical. Hardness values averaged for 25 nanoindentation curves of each sample for indentation depth from 2.0 to 2.5 μm are 33.5 (±0.74) MPa for ball-milled, 31.1 (±1.79) MPa for prestrained, and 32.1 (±0.79) MPa for annealed np-Au samples; again, the values are almost identical. These results indicate that grain size and initial dislocation density are not critical in determining the nanoindentation hardness of np-Au of ligament size about 70 nm, while nanoindentation hardness of np-Au is known to be strongly dependent on ligament size.16,18 The compressive strength of a prestrained Au nanopillar with diameter of 200−300 nm is lower than that of a pristine one since prestraining introduces weak dislocation sources such as mobile dislocations and singlearm sources.27 However, the high initial dislocation density of np-Au with ligament size about 70 nm does not affect the nanoindentation hardness. Unlike the uniaxial compression of the nanopillar, ligaments, and junctions in np-Au experience such a complicated stress state in nanoindentation that structural stress concentrators are likely to be more critical factors than dislocation sources inside ligaments. Figure 3a−c shows SEM images of nanoindentation marks on the top of np-Au samples. A common feature of these three images is that ligaments beneath nanoindenter do not experience severe deformation during nanoindentation. We measured relative density, the ratio of solid ligaments to entire volume of np-Au, by measuring the weight and dimension of

Figure 2. Nanoindentations on precursor alloys and np-Au samples. (a) Typical and (b) averaged force versus indentation depth curves of precursor alloys. A greater force is required to attain maximum indentation depth of 2.5 μm for (in order) ball-milled, prestrained, and annealed precursor alloys. (c) Typical and (d) averaged force versus indentation depth curves for three np-Au samples that almost overlap.

attained at maximum indentation depth of 2.5 μm are in the order of ball-milled, prestrained, and annealed precursor alloys. Hardness averaged for 25 nanoindentation curves of each sample for indentation depth from 2 to 2.5 μm is 0.93 (±0.04) GPa for ball-milled, 0.63 (±0.06) GPa for prestrained, and 0.48 (±0.01) GPa for annealed samples, while elastic moduli in this range are similar: 76.0 (±2.8) GPa for ball-milled, 78.1 (±2.6) GPa for prestrained, and 72.6 (±2.0) GPa for annealed samples. The greatest hardness of the ball-milled precursor alloy is attributed to refinement of grain size and extremely high initial dislocation density because of the severe plastic

Figure 3. SEM images of indentation marks after unloading for three np-Au samples. (a−c) Low-magnification SEM images for annealed, prestrained, and ball-milled, respectively. (d−f) High-magnification SEM images. Grain boundary sliding by nanoindentation, as indicated by arrows in f, is observed in ball-milled np-Au. C

DOI: 10.1021/acs.nanolett.6b00062 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters np-Au samples. The relative density of np-Au samples in this study was 28%. Even ligaments directly beneath the nanoindenter are likely to experience an increase in relative density by structural plastic collapse rather than by severe deformation of ligaments. While this behavior agrees with a previous report for np-Au with a relative density of 25%, ligaments of np-Au with relative density 42% beneath nanoindenter show severe deformations.16,39 We believe that the relative density of np-Au could be a primary factor in this difference. As in the Gibson− Ashby model, yield strength and elastic modulus increase with increasing relative density.15 This indicates that, for np-Au of higher relative density, undeformed region can support collapsed region more strongly; thus the collapsed region can be densified by severe deformation of ligaments. For lower relative density, collapsed region is more likely to expand to the undeformed region rather than creating additional densification of the collapsed region by severe deformation of ligaments. Figure 3d−f shows high-magnification SEM images of the nanoindentation marks. Grain boundary sliding is observed in the ball-milled np-Au (indicated by arrow in Figure 3f; see Supporting Information, Figure S2), and Figures 3d and e show typical plastic collapse features. Reports on compression of bicrystalline pillars show that a vertically aligned grain boundary increases strength compared to single-crystalline pillars,25,40 whereas a grain boundary inclined to the loading direction slides in shear stress about 40% below the flow stress of bicrystalline pillar including a vertical grain boundary, which is similar to the flow stress of single crystalline pillars.41 Grain boundary density is very high in ball-milled np-Au, and their distribution is random. Some grain boundaries are seen to slide during nanoindentation, but this grain boundary sliding does not affect nanoindentation hardness significantly. Figure 4a shows typical flexural stress−strain curves for three np-Au samples in three-point bending test, indicating that npAu samples show brittle fracture right after the linear elastic regime. Flexural stress σf is given by σf = 3PL/2wt2, where P is force, L outer span length, w sample width, and t sample thickness. Flexural strain, εf is expressed by εf = 6vt/L2, where v is the displacement. Flexural strength is 8.16 (±1.47) MPa for annealed np-Au, 9.03 (±2.24) MPa for prestrained np-Au, and 5.41 (±0.68) MPa for ball-milled np-Au (Figure 4b). The flexural strength of ball-milled np-Au is 33.7% lower than that of annealed np-Au and 40.1% lower than that of prestrained npAu. The high density of grain boundaries in ball-milled np-Au weakens flexural strength. To investigate this phenomenon, we examined fracture surfaces using SEM as presented in Figure 4c−h. Figure 4c and f show fracture surface for the highest flexural strength of 12.51 MPa among several data for prestrained np-Au. The fracture surface is flat at low magnification, and the ligaments show ductile fracture features, necking and point-to-point separation. This is typical transgranular fracture in np-Au.42,43 Figure 4d and g shows the fracture surface for the lowest flexural strength, 5.55 MPa, among the flexural strengths of annealed and prestrained npAu. This fracture surface shows a mixed configuration of transgranular and intergranular fracture. A high-magnification image of the intergranular fracture surface shows area-to-area separations of ligaments (see Figure 4g), unlike the point-topoint separation by transgranular fracture in Figure 4f. During dealloying, Ag atoms at grain boundaries leach faster than those in the grain interior, so that grain boundaries in npAu contain excess volumes.42 Furthermore, Ag atoms segregate to interfaces such as surface or grain boundaries during

Figure 4. Mechanical properties and fracture surfaces obtained by three-point bending tests of three np-Au samples. (a) Typical flexural stress−strain curves for three np-Au samples that show brittle fracture right after linear elastic deformation. (b) Flexural strength of three npAu samples showing that flexural strength of ball-milled np-Au is significantly lower than those of annealed and prestrained np-Au, and scatter in the results is also smaller than in annealed and prestrained np-Au. (c, f) SEM images of fracture surface formed by transgranular fracture for highest flexural strength of annealed np-Au. (d, g) Fracture surface formed by mixture of transgranular and intergranular fracture for lowest flexural strength of prestrained np-Au. (e, h) Fracture surface formed by intergranular fracture for ball-milled np-Au. Greatest tensile stress was applied during three-point bending test at sample bottoms in (c−e). Scale bar of inset images in f−h is 500 nm.

annealing of the precursor alloys.44 The amount of residual Ag was about 2 at.% for the three np-Au samples in this study. Excess volume and segregated Ag atoms at grain boundaries induce localized stress, and thus grain boundaries act as twodimensional void-like defects. It seems reasonable from previous experiments that grain boundaries weaken np-Au, especially in tensile or flexural testing. Li and Sieradzki measured fracture strength and strain of np-Au in three-point bending tests.11 The grain size of np-Au in their work was not reported but could be assumed to be on order of 100 μm or greater, since they made np-Au from precursor alloys annealed at 900 °C for 24 h. Calculated fracture strengths are 8.2 MPa for ligament size 18 nm and 12 MPa for ligament size 60 nm, values similar to our flexural strengths. Biener et al. performed three-point bending tests to compare fracture behaviors of npAu with different ligament sizes.42 They found that the strength of np-Au depends on “the largest critical defects” such as grainboundary opening regions where stress is concentrated. Balk et al. compared tensile and compressive behavior of polycrystalline np-Au microsample.45 Fractography of np-Au showed that individual ligaments undergo ductile failure by necking even though np-Au exhibits macroscopic brittle fracture. Furthermore, the ultimate compressive strength is twice the tensile fracture strength. Briot et al., comparing the tensile properties of polycrystalline and single-crystalline np-Au using atomistic simulations, found that the strength of single-crystalline np-Au D

DOI: 10.1021/acs.nanolett.6b00062 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Excess volume by faster dealloying at grain boundaries than inside grains, segregated Ag atoms at grain boundaries, and reduction of cross-sectional area by cusp formation at grain boundaries were suggested as possible mechanisms for the grain boundary-induced weakening in the flexural strength of np-Au.

is three times greater than that of polycrystalline np-Au, while cracks initiate at grain boundaries with low connectivity.46 When grain boundaries parallel to the span direction in threepoint bending are present in the middle of the outer surface where maximum tensile stress is applied, fracture initiates by grain-boundary opening, leading to low flexural strength. Otherwise, high flexural strength is attained by transgranular fracture. Due to the relatively large grain size of annealed and prestrained np-Au, flexural strength seems to depend on the stochastic distribution of grain boundaries, explaining the larger scatter in flexural strength for annealed and prestrained np-Au than for ball-milled np-Au. Figure 4e and h reveal that the fracture surface was formed by uniform intergranular fracture, possibly due to the small grain size (206 nm) in ball-milled npAu. This clearly explains the much lower flexural strength and scatter for ball-milled np-Au. Zhong et al. prepared np-Au samples from four Au−Cu precursor alloys: as-cast, annealed, cold-rolled, high-pressure torsion (HPT) ones.47 They found intergranular cleavage at the fracture surface of annealed np-Au, although Cu (unlike Ag) does not segregate at grain boundaries, which is attributed to preferential dealloying due to higher energy state at grain boundaries. For np-Au made from Au−Ag precursor alloys, segregation of residual Ag atoms at grain boundaries during annealing can weaken grain boundaries. Since ball-milling can randomize the segregated Ag atoms, weakening of the grain boundaries by segregation of Ag atoms in ball-milled np-Au could be less than that in annealed np-Au. They also found that grain boundaries with higher excess volume can be formed in cold-rolled and HPTtreated np-Au due to irregular local stress induced during severe deformation of precursor alloys. This could also be a possible weakening mechanism of grain boundaries in ball-milled np-Au. Additionally, cusps at grain boundaries in ligaments, as seen in Figure 1l, can act as stress concentrators due to the reduction in cross-sectional area along with grain boundary and this, in addition to the possible mechanisms mentioned above, could make grain boundary opening easier. The mechanical behavior of np-Au with different densities of initial dislocations and grain boundaries has been investigated by nanoindentation and three-point bending test. Nanocrystalline np-Au was fabricated by free corrosion dealloying from ball-milled precursor alloy, and prestrained np-Au was fabricated from prestrained precursor in 5% compressive engineering strain. Contrary to typical behavior in solid materials, high densities of grain boundaries and initial dislocations did not strengthen np-Au. In nanoindentation on np-Au samples, the hardness of ball-milled and prestrained npAu is neither lower nor higher than that annealed np-Au, while ligaments of ball-milled np-Au have shown sheared-off structures along with grain boundaries. In three-point bending tests, np-Au samples have shown catastrophic brittle fracture, and the flexural strengths were 8.16 MPa for annealed np-Au, 9.03 MPa for prestrained np-Au and 5.41 MPa for ball-milled np-Au, a phenomenon called an grain boundary-induced weakening of flexural strength in np-Au. The relatively large scatter in flexural strength for annealed and prestrained np-Au samples depends stochastically on the distribution of grain boundaries acting as two-dimensional defects with low connectivity. Ball-milled np-Au has a much greater density of two-dimensional defects than annealed and prestrained np-Au, where intergranular fracture is preferred. Therefore, the probable existence of grain boundary opening in the highest tensile region is attributed to the flexural strength of np-Au.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.6b00062. Experimental details of sample preparation and mechanical testing, calculation of initial dislocation density of annealed and prestrained precursors, and SEM and EBSD IPF image of ligaments deformed by nanoindentation (PDF) Supporting movie of three-point bending test (AVI)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science, ICT & Future Planning (MSIP) (No. NRF-2015R1A5A1037627), and by the KIST-UNIST partnership program (1.150091.01/ 2.150464.01).

■

REFERENCES

(1) Wittstock, A.; Zielasek, V.; Biener, J.; Friend, C. M.; Baumer, M. Science 2010, 327 (5963), 319−322. (2) Biener, M. M.; Biener, J.; Wichmann, A.; Wittstock, A.; Baumann, T. F.; Baumer, M.; Hamza, A. V. Nano Lett. 2011, 11 (8), 3085−3090. (3) Fujita, T.; Guan, P. F.; McKenna, K.; Lang, X. Y.; Hirata, A.; Zhang, L.; Tokunaga, T.; Arai, S.; Yamamoto, Y.; Tanaka, N.; Ishikawa, Y.; Asao, N.; Yamamoto, Y.; Erlebacher, J.; Chen, M. W. Nat. Mater. 2012, 11 (9), 775−780. (4) Hu, K. C.; Lan, D. X.; Li, X. M.; Zhang, S. S. Anal. Chem. 2008, 80 (23), 9124−9130. (5) Jia, F. L.; Yu, C. F.; Ai, Z. H.; Zhang, L. Z. Chem. Mater. 2007, 19 (15), 3648−3653. (6) Biener, J.; Wittstock, A.; Zepeda-Ruiz, L. A.; Biener, M. M.; Zielasek, V.; Kramer, D.; Viswanath, R. N.; Weissmuller, J.; Baumer, M.; Hamza, A. V. Nat. Mater. 2009, 8 (1), 47−51. (7) Kramer, D.; Viswanath, R. N.; Weissmuller, J. Nano Lett. 2004, 4 (5), 793−796. (8) Detsi, E.; Punzhin, S.; Rao, J. C.; Onck, P. R.; De Hosson, J. T. M. ACS Nano 2012, 6 (5), 3734−3744. (9) Erlebacher, J.; Aziz, M. J.; Karma, A.; Dimitrov, N.; Sieradzki, K. Nature 2001, 410 (6827), 450−453. (10) Ding, Y.; Kim, Y. J.; Erlebacher, J. Adv. Mater. 2004, 16 (21), 1897. (11) Li, R.; Sieradzki, K. Phys. Rev. Lett. 1992, 68 (8), 1168−1171. (12) Parida, S.; Kramer, D.; Volkert, C. A.; Rosner, H.; Erlebacher, J.; Weissmuller, J. Phys. Rev. Lett. 2006, 97 (3), 035504. (13) Lee, D.; Wei, X.; Chen, X.; Zhao, M.; Jun, S. C.; Hone, J.; Herbert, E. G.; Oliver, W. C.; Kysar, J. W. Scr. Mater. 2007, 56 (5), 437−440. (14) Volkert, C. A.; Lilleodden, E. T.; Kramer, D.; Weissmuller, J. Appl. Phys. Lett. 2006, 89 (6), 3. (15) Gibson, L. J.; Ashby, M. F. Proc. R. Soc. London, Ser. A 1982, 382 (1782), 43. E

DOI: 10.1021/acs.nanolett.6b00062 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters (16) Biener, J.; Hodge, A. M.; Hayes, J. R.; Volkert, C. A.; ZepedaRuiz, L. A.; Hamza, A. V.; Abraham, F. F. Nano Lett. 2006, 6 (10), 2379−2382. (17) Sun, Y.; Ye, J.; Shan, Z.; Minor, A. M.; Balk, T. J. JOM 2007, 59 (9), 54−58. (18) Hodge, A. M.; Biener, J.; Hayes, J. R.; Bythrow, P. M.; Volkert, C. A.; Hamza, A. V. Acta Mater. 2007, 55 (4), 1343−1349. (19) Farkas, D.; Caro, A.; Bringa, E.; Crowson, D. Acta Mater. 2013, 61 (9), 3249−3256. (20) Greer, J. R.; Oliver, W. C.; Nix, W. D. Acta Mater. 2005, 53 (6), 1821−1830. (21) Uchic, M. D.; Dimiduk, D. M.; Florando, J. N.; Nix, W. D. Science 2004, 305 (5686), 986−989. (22) Greer, J. R.; Nix, W. D. Phys. Rev. B: Condens. Matter Mater. Phys. 2006, 73 (24), 6. (23) Greer, J. R.; De Hosson, J. T. M. Prog. Mater. Sci. 2011, 56 (6), 654−724. (24) Volkert, C. A.; Lilleodden, E. T. Philos. Mag. 2006, 86 (33−35), 5567−5579. (25) Ng, K. S.; Ngan, A. H. W. Philos. Mag. 2009, 89 (33), 3013− 3026. (26) Kunz, A.; Pathak, S.; Greer, J. R. Acta Mater. 2011, 59 (11), 4416−4424. (27) Lee, S. W.; Han, S. M.; Nix, W. D. Acta Mater. 2009, 57 (15), 4404−4415. (28) Bei, H.; Shim, S.; Pharr, G. M.; George, E. P. Acta Mater. 2008, 56 (17), 4762−4770. (29) ASTM C1674−08, Standard Test Method for Flexural Strength of Advanced Ceramics with Engineered Porosity (Honeycomb Cellular Channels) at Ambient Temperatures; ASTM International: West Conshohocken, PA, 2008. (30) Jin, H. J.; Kurmanaeva, L.; Schmauch, J.; Rosner, H.; Ivanisenko, Y.; Weissmuller, J. Acta Mater. 2009, 57 (9), 2665−2672. (31) Van Petegem, S.; Brandstetter, S.; Maass, R.; Hodge, A. M.; ElDasher, B. S.; Biener, J.; Schmitt, B.; Borca, C.; Van Swygenhoven, H. Nano Lett. 2009, 9 (3), 1158−1163. (32) Dotzler, C. J.; Ingham, B.; Illy, B. N.; Wallwork, K.; Ryan, M. P.; Toney, M. F. Adv. Funct. Mater. 2011, 21 (20), 3938−3946. (33) Doña, J. M.; González-Velasco, J. J. Phys. Chem. 1993, 97 (18), 4714−4719. (34) Lu, X.; Balk, T. J.; Spolenak, R.; Arzt, E. Thin Solid Films 2007, 515 (18), 7122−7126. (35) Gwak, E. J.; Kang, N. R.; Baek, U. B.; Lee, H. M.; Nahm, S. H.; Kim, J. Y. Scr. Mater. 2013, 69 (10), 720−723. (36) Kertis, F.; Snyder, J.; Govada, L.; Khurshid, S.; Chayen, N.; Erlebacher, J. JOM 2010, 62 (6), 50−56. (37) Sun, Y.; Burger, S. A.; Balk, T. J. Philos. Mag. 2014, 94 (10), 1001−1011. (38) Nix, W. D.; Gao, H. J. J. Mech. Phys. Solids 1998, 46 (3), 411− 425. (39) Biener, J.; Hodge, A. M.; Hamza, A. V.; Hsiung, L. M.; Satcher, J. H. J. Appl. Phys. 2005, 97 (2), 024301. (40) Kheradmand, N.; Vehoff, H. Adv. Eng. Mater. 2012, 14 (3), 153−161. (41) Aitken, Z. H.; Jang, D. C.; Weinberger, C. R.; Greer, J. R. Small 2014, 10 (1), 100−108. (42) Biener, J.; Hodge, A. M.; Hamza, A. V. Appl. Phys. Lett. 2005, 87 (12), 121908. (43) Sun, S. F.; Chen, X. Y.; Badwe, N.; Sieradzki, K. Nat. Mater. 2015, 14 (9), 894. (44) Meinel, K.; Klaua, M.; Bethge, H. Phys. Status Solidi A-Appl. Res. 1988, 106 (1), 133−144. (45) Balk, T. J.; Eberl, C.; Sun, Y.; Hemker, K. J.; Gianola, D. S. JOM 2009, 61 (12), 26−31. (46) Briot, N. J.; Kennerknecht, T.; Eberl, C.; Balk, T. J. Philos. Mag. 2014, 94 (8), 847−866. (47) Zhong, Y.; Markmann, J.; Jin, H. J.; Ivanisenko, Y.; Kurmanaeva, L.; Weissmuller, J. Adv. Eng. Mater. 2014, 16 (4), 389−398.

F

DOI: 10.1021/acs.nanolett.6b00062 Nano Lett. XXXX, XXX, XXX−XXX